An modified PBIL for network coding resource minimization in dynamic network environment

In network coding, intermediate nodes are allowed to mathematically recombine packets received from different incoming links, which helps increase network throughput and accommodate more traffic flows with limited network resources. Coding operations (i.e. packet recombination), however, could cause significant computational cost and thus introduce heavy burden to the network if they are performed wherever possible. It is hence important to always keep the amount of coding operations minimized in a dynamic network environment. This paper proposes a modified population based incremental learning (PBIL) for solving the above problem, where an environmental adaptation scheme is devised to guide the search tracing the ever-changing optima within the fitness landscape in a dynamic network environment. Experimental results show that the proposed PBIL gains better performance than several state-of-the-art evolutionary algorithms regarding the solution quality.